OpenAI 召集了约 40 位数学家讨论人工智慧超越人类能力时的应对措施，并透露其模型已解决上百个长期未决的数学难题，甚至包含千禧年大奖难题。然而，数学界要求以严谨学术论文发表成果的建议遭到忽视，OpenAI 仍计划直接在 GitHub 发布大量解答，引发学术圈对其缺乏同行评审与严谨公开程序的忧虑与不满。

在与 Anthropic 争夺技术领先与筹备上市的竞争中，OpenAI 被指责摒弃传统科学规范，甚至引发抢夺学术成果与作者署名的争端。纽约大学教授 Tristan Buckmaster 控诉该公司抢先发表其与 Anthropic 研究员合作未公开的成果，并在协商过程中遭受威胁，致使学者们对科技巨头权力过度集中及「黑帮式行为」感到高度焦虑与反感。

尽管部分 OpenAI 内部人员认为 AI 将终结数学家职业，多数数学家仍拒绝承认数学已死，并尝试建立如 Hexagon 和 Palomar 等新工具来吸收验证 AI 生成的证明。数学界普遍认为这项强大工具改变了研究模式，要求科技公司改善透明度与揭露机制，以兼具恐惧与兴奋的心态迎向这波转型。

OpenAI convened around 40 mathematicians to address the implications of AI surpassing human mathematical abilities, revealing that its models had resolved over 100 open problems, including a Millennium Prize problem. However, the company largely disregarded the academic community's pleas for rigorous research papers, opting instead to dump hundreds of solutions directly onto GitHub without standard scientific dissemination.

Driven by fierce competition with Anthropic ahead of anticipated IPOs, AI labs have been criticized for casting aside established scientific practices, leading to intense disputes over attribution and credit. Tristan Buckmaster, a mathematics professor at NYU, accused OpenAI of front-running his unpublished collaborative work and facing coercive behavior during authorship disputes, heightening scholars' concerns over unchecked corporate power.

While some inside OpenAI believe the rapid rise of AI models will render traditional math careers obsolete, mathematicians reject the notion that their field is dead, instead establishing platforms like Hexagon and Palomar to verify machine-generated proofs. Academics view these tools as transformative rather than terminal, emphasizing that corporate transparency and responsible disclosure are vital during this profoundly unsettling yet exciting era.